The discovery of BRCA1 and BRCA2 has resulted in more appropriate targeting of preventive and screening strategies for breast cancer. The on-going discovery of genes with moderate-risk mutations, and of multiple loci with common low-risk associated variants, means that more women at substantial genetic risk will be identified. Despite these advances in genomic medicine, there remain major unanswered questions for high risk women, the majority of whom do not carry mutations in any currently identified susceptibility genes: 1) What is my absolute risk of breast cancer?; 2) Are there modifiable factors that might lower my risk?; and, for women with prior breast cancer, 3) Can I do anything to lower my risk of a new cancer? Answers to these questions are fundamental to improving clinical care, and are long overdue. We lack answers to these important questions because many studies fail to capture the complexity of family history and lack long-term follow-up data to measure risk. Breast cancer risk prediction models commonly used at non-specialist clinics often capture risk based on only first-degree family history. No breast cancer prediction models have been based on, nor validated with, large prospective cohorts of high risk women. Studies that have examined potential modifiers of risk for BRCA1 and BRCA2 mutation carriers have used a retrospective design and included prevalent cancers over-sampled for disease survivors. To address these gaps, we propose to conduct active follow-up of 30,563 women of whom 2,597 are BRCA1 and BRCA2 mutation carriers. These women come from 9,739 families recruited and followed since 1995 in the U.S., Canada, and Australia. We collected the same extensive baseline epidemiologic, multigenerational pedigree, and genetic data for these women. Our prospective family study is enriched with women at increased susceptibility for breast cancer who vary widely in underlying Familial Risk Profile (FRP), which can be estimated using multigenerational pedigree and genetic data. We will estimate age-specific absolute, and relative, risks of breast cancer using two separate cohorts (18,530 women unaffected and 12,033 women affected at baseline), as a function of their estimated FRP, modifiable risk factors, and by BRCA1 and BRCA2 mutation status. By the end of follow-up, we estimate 1,427 of the women unaffected and 1,359 women affected at baseline will be diagnosed with a new breast cancer. 15-17% of these new cases will be in BRCA1 or BRCA2 mutation carriers. We will use our findings to enhance prediction models by incorporating information from multigenerational family history, measured gene variants, and risk factors. Clinical practice has been conservative in advising high risk women, particularly mutation carriers, about potential lifestyle modifications to reduce risk, basing this advice on studies of average-risk women. Instead, we propose to build more accurate prediction models for women across the spectrum of risk that can be used to tailor more effective prevention strategies.